Filters








445 Hits in 4.5 sec

Just-in-Time Learning for Bottom-Up Enumerative Synthesis [article]

Shraddha Barke, Hila Peleg, Nadia Polikarpova
2020 pre-print
In this work we introduce an alternative approach to guided program synthesis: instead of training a model ahead of time we show how to bootstrap one just in time, during synthesis, by learning from partial  ...  In response to this challenge, recent work proposed to guide synthesis using learned probabilistic models.  ...  ACKNOWLEDGMENTS The authors would like to thank the anonymous reviewers for their feedback on the draft of this paper.  ... 
doi:10.1145/3428295 arXiv:2010.08663v1 fatcat:f6ldtg4b2bcq3khujo223f2fha

Predictive Synthesis of API-Centric Code [article]

Daye Nam, Baishakhi Ray, Seohyun Kim, Xianshan Qu, Satish Chandra
2022 arXiv   pre-print
In this work, we examine ways in which machine learning can be used to accelerate enumerative program synthesis.  ...  DeepCoder) in which researchers have used ML models in enumerative synthesis.  ...  Conclusion In this paper, we proposed a new machine learning technique to speed up enumerative program synthesis.  ... 
arXiv:2201.03758v2 fatcat:x5ozkupjizdtzkvc5mpcfjea7a

Just-in-Time Learning for Inductive Program Synthesis Artifact [article]

Shraddha Barke, Hila Peleg, Nadia Polikarpova
2020 Zenodo  
The paper's title has since then been changed to "Just-in-Time Learning for Bottom-up Enumerative Synthesis".  ...  This code is the implementation for the first version of the paper titled "Just-in-Time Learning for Inductive Program Synthesis" submitted to OOPSLA 2020 AEC.  ...  Experiments from Section 6.2 This section covers the experiments from Section 6.2 of the paper that evaluates the effectiveness of the just-in-time learning approach.  ... 
doi:10.5281/zenodo.4039224 fatcat:3diriolgr5gpjmd7z3towz2h7i

TF-Coder: Program Synthesis for Tensor Manipulations [article]

Kensen Shi, David Bieber, Rishabh Singh
2020 arXiv   pre-print
TF-Coder uses a bottom-up weighted enumerative search, with value-based pruning of equivalent expressions and flexible type- and value-based filtering to ensure that expressions adhere to various requirements  ...  In this work, we present a tool called TF-Coder for programming by example in TensorFlow.  ...  ACKNOWLEDGMENTS The authors thank Charles Sutton and the other members of the program synthesis team at Google Brain for helpful discussions.  ... 
arXiv:2003.09040v3 fatcat:cvgj7xrshfhxhhqjpsyv7nyjf4

TF-Coder: Program Synthesis for Tensor Manipulations

Kensen Shi, David Bieber, Rishabh Singh
2022 ACM Transactions on Programming Languages and Systems  
TF-Coder uses a bottom-up weighted enumerative search, with value-based pruning of equivalent expressions and flexible type- and value-based filtering to ensure that expressions adhere to various requirements  ...  In this work, we present a tool called TF-Coder for programming by example in TensorFlow.  ...  ACKNOWLEDGMENTS The authors thank Charles Sutton and the other members of the program synthesis team at Google Brain for helpful discussions.  ... 
doi:10.1145/3517034 fatcat:dqqhqvtevnefdgn7zihokv636u

One Down, 699 to Go: or, synthesising compositional desugarings [article]

Sándor Bartha and James Cheney and Vaishak Belle
2021 arXiv   pre-print
We evaluate enumerative synthesis as a baseline algorithm, and demonstrate that, with our reformulation of the problem, it is possible to learn correct desugaring rules for the example source and core  ...  s desugaring learning framework in order to clarify the assumptions necessary for an incremental learning algorithm to be feasible.  ...  ACKNOWLEDGEMENTS We would like to thank the anonymous reviewers and our shepherd Shriram Krishnamurthi for helpful feedback and suggestions for improvement.  ... 
arXiv:2109.06114v1 fatcat:hja4ks7gp5aehlmw4kyau2p4fm

Neural-Guided Deductive Search for Real-Time Program Synthesis from Examples [article]

Ashwin Kalyan, Abhishek Mohta, Oleksandr Polozov, Dhruv Batra, Prateek Jain, Sumit Gulwani
2018 arXiv   pre-print
to provide real-time synthesis on challenging benchmarks.  ...  In this work, we propose Neural Guided Deductive Search (NGDS), a hybrid synthesis technique that combines the best of both symbolic logic techniques and statistical models.  ...  Most approaches employ either bottom-up enumerative search (Udupa et al., 2013) , constraint solving (Torlak & Bodik, 2013) , or inductive logic programming (Lin et al., 2014) , and thus scale poorly  ... 
arXiv:1804.01186v2 fatcat:bdovzu7cfrcrda2n4umckjs3na

Scaling Neural Program Synthesis with Distribution-based Search [article]

Nathanaël Fijalkow and Guillaume Lagarde and Théo Matricon and Kevin Ellis and Pierre Ohlmann and Akarsh Potta
2021 arXiv   pre-print
Collectively these findings offer theoretical and applied studies of search algorithms for program synthesis that integrate with recent developments in machine-learned program synthesizers.  ...  Within this framework, we introduce two new search algorithms: Heap Search, an enumerative method, and SQRT Sampling, a probabilistic method.  ...  We present a new efficient and loss optimal algorithm called Heap Search and following a bottom-up approach.  ... 
arXiv:2110.12485v1 fatcat:2a6rudsktraohmeni7ehsljs5i

Bottom-up Synthesis of Recursive Functional Programs using Angelic Execution [article]

Anders Miltner and Adrian Trejo Nuñez and Ana Brendel and Swarat Chaudhuri and Isil Dillig
2021 arXiv   pre-print
While bottom-up synthesis techniques can work better than top-down methods in certain settings, there is no prior technique for synthesizing recursive programs from logical specifications in a purely bottom-up  ...  We present a novel bottom-up method for the synthesis of functional recursive programs.  ...  For a broader survey of program synthesis, see Gulwani et al. [2017]. Bottom-up Synthesis. Bottom-up enumeration is a classic approach to program synthesis.  ... 
arXiv:2107.06253v2 fatcat:36h2qvwupbecrpyy7w3cys4pjy

Quantitative Programming by Examples [article]

Sumit Gulwani and Kunal Pathak and Arjun Radhakrishna and Ashish Tiwari and Abhishek Udupa
2019 arXiv   pre-print
Our detailed experiments validate the design of our procedure and show the value of combining global and local search for qPBE.  ...  We present a modular approach for solving qPBE that consists of three phases: intent disambiguation, global search, and local search.  ...  Some PBE engines that support enumerative search (bottom up synthesis) can be adapted to solve the local qPBE synthesis problem.  ... 
arXiv:1909.05964v1 fatcat:lg53xflw2jbnpajaxekm7okv7u

Programming by Examples: PL Meets ML [chapter]

Sumit Gulwani, Prateek Jain
2017 Lecture Notes in Computer Science  
We make the case for synthesizing these heuristics from training data using appropriate machine learning methods.  ...  There are three key components in a PBE system. (i) A search algorithm that can efficiently search for programs that are consistent with the examples provided by the user.  ...  in this article related to using ML techniques for search and ranking.  ... 
doi:10.1007/978-3-319-71237-6_1 fatcat:nou2fnkpt5elfj3ohaunnfmy7y

Synthesizing highly expressive SQL queries from input-output examples

Chenglong Wang, Alvin Cheung, Rastislav Bodik
2017 Proceedings of the 38th ACM SIGPLAN Conference on Programming Language Design and Implementation - PLDI 2017  
Our results showed that SCYTHE efficiently solved 74% of the benchmarks, most in just a few seconds.  ...  Using abstract queries to represent the search space nicely decomposes the synthesis problem into two tasks: (1) searching for abstract queries that can potentially satisfy the given I/O examples, and  ...  Acknowledgements This work is supported in part by the National Science Foundation through grants IIS-1546083, IIS-1651489, and CNS-1563788; DARPA award FA8750-16-2-0032; DOE award DE-SC0016260; and gifts  ... 
doi:10.1145/3062341.3062365 dblp:conf/pldi/WangCB17 fatcat:ap2kzhjks5cinofsafl3xxs3oe

Top-Down oder Bottom-Up?/Top-Down or Bottom-Up?

Michael Fritsch
1992 Jahrbücher für Nationalökonomie und Statistik  
In this work we present a novel program synthesis approach which combines the benefits of deductive and enumerative synthesis strategies, yielding a semi-supervised technique with which concise programs  ...  Automatic synthesis of web data extraction programs has been explored in a variety of settings, but in practice there remain various robustness and usability challenges.  ...  Although the bottom-up synthesis is mainly used for an unsupervised analysis of the webpage, in the final step of the main algorithm ( Figure 5 ) we resort to a purely bottom-up search if no satisfying  ... 
doi:10.1515/jbnst-1992-5-604 fatcat:mmztkyctvfgzvozq6jzrgaflqm

Learning to Represent Programs with Property Signatures [article]

Augustus Odena, Charles Sutton
2020 arXiv   pre-print
We introduce the notion of property signatures, a representation for programs and program specifications meant for consumption by machine learning algorithms.  ...  one-tenth of the time.  ...  ACKNOWLEDGMENTS We would like to thank Kensen Shi, David Bieber, and the rest of the Program Synthesis Team for helpful discussions.  ... 
arXiv:2002.09030v1 fatcat:ew5evgqhtjcwdllz4oc5fj75x4

Automated transpilation of imperative to functional code using neural-guided program synthesis

Benjamin Mariano, Yanju Chen, Yu Feng, Greg Durrett, Işil Dillig
2022 Proceedings of the ACM on Programming Languages (PACMPL)  
Motivated by this problem, this paper presents a transpilation approach based on inductive program synthesis for modernizing existing code.  ...  identical low-level expressions, but these expressions also take the same values in corresponding execution traces.  ...  ACKNOWLEDGMENTS We would like to thank Benjamin Sepanski, Shankara Pailoor, and Jocelyn Chen for their thoughtful feedback.  ... 
doi:10.1145/3527315 fatcat:awo63jde3jfd5hhj7uukucdqja
« Previous Showing results 1 — 15 out of 445 results